How to Enable Agent-to-Agent Speech Detection in LiveKit

By default, LiveKit agents cannot detect speech from other agents in the same room. To enable agent-to-agent conversations where agents can hear and respond to each other, you'll need to configure the RoomInputOptions.

Configuring Agent Speech Detection

Add the following configuration to enable speech detection between agents:

RoomInputOptions.participant_kinds = [rtc.ParticipantKind.PARTICIPANT_KIND_AGENT]

This setting allows agents to:

• Detect speech from other agents in the same room

• Process other agents' speech as input

• Maintain two-way conversations with other agents